Methodological approach to assess biodiversity in agro ecosystems – A case study Abordagem metodológica para a avaliação da biodiversidade em ecossistemas agrários - estudo de caso Catarina Tavares1, Ana F. Gouveia2, Luís Crespo3, Célia Mateus4 e M. Teresa Rebelo5
1 Departamento de Recursos Naturais, Ambiente e Território. Tapada da Ajuda, 1349-017 Lisboa, Portugal; [email protected], author for correspondence 2 AmBioDiv - Valor Natural, Lda. Rua Filipe da Mata, 10, 1º Frente, 1600-071 Lisboa, Portugal 3 IMAR –Universidade dos Açores, Grupo da Biodiversidade dos Açores (CITA-A). 9700-042 Angra do Heroísmo Terceira, Açores, Portugal 4 Instituto Nacional de Investigação Agrária e Veterinária, IP. Av. República, Quinta do Marquês, 2784-505 Oeiras, Portugal 5 Departamento de Biologia Animal/ Centro de Estudos do Ambiente e do Mar (CESAM), Faculdade de Ciências da Universidade de Lisboa. Rua Ernesto Vasconcelos, Ed. C2, Campo Grande, 1749-016 Lisboa, Portugal
Recebido/Received: 2012.07.12 Aceitação/Accepted: 2013.05.10
ABSTRACT In agroecosystems, the evaluation of biodiversity is extremely important given its relation with their sustainability, how- ever it is a difficult and expensive task. The use of the higher taxa surrogacy has been proposed as an effective and less expensive way to assess biodiversity, but is still being used mainly in natural, more stable, ecosystems. In this study, we tested the use of the higher taxa surrogacy approach for spiders in pear orchards, under different production systems, in the “Oeste” region of Portugal. Spiders were chosen for this case study, since they are important predators, acting as biological control agents in agroecosystems. We concluded that genus can be used as a higher taxa surrogate of species richness with a good accuracy. Key-Words: Higher taxa surrogacy, pear orchards, spiders.
RESUMO Em ecossistemas agrícolas, a avaliação da biodiversidade é extremamente importante dada a sua relação com a sustentabilidade, sendo contudo uma tarefa difícil e dispendiosa. O uso da relação entre o número de espécies e o número de táxones de alto nível, tem sido proposto como uma alternativa eficaz e económica para a avaliação da biodiversidade, contudo a sua utilização tem estado praticamente restringida aos ecossistemas naturais, mais estáveis. Neste estudo, testou-se o uso da relação entre o número de espécies e o número de táxones de alto nível em aranhas, em pomares de pêra sob diferentes modos de produção, na região Oeste de Portugal. As aranhas foram seleccionadas para este estudo de caso por serem importantes predadores, actuando como agentes de luta biológica contra pragas. Concluiu-se que o género pode ser usado na avaliação da riqueza de espécies, com um elevado grau de precisão. Palavras-chave: Aranhas, pomares de pêra, relação entre táxones
Introduction sity (Moonen and Bàrberi, 2008; Paoletti, 1999). Bio- diversity is often measured by the number of species In the last few years, the interest in associating na- sampled in an ecosystem (Primack, 2002), which is a ture conservation to agriculture has increased enor- rather difficult and expensive task, especially when mously in Europe and several measures have been considering invertebrate taxa (Duelli et al., 1999; Mar- taken for the implementation of a sustainable agri- tín-Piera, 2000). However, since invertebrates account culture (Amaro, 2003). for more than 90% of the genetic variability in agroe- One way of measuring the sustainability of an agro- cosystems, it seems necessary to use them as the clos- ecosystem is through the evaluation of its biodiver- est correlates to the overall biodiversity (Duelli, 1997).
324 Revista de Ciências Agrárias, 2013, 36(3): 324-330 The use of the higher taxa surrogacy approach (or- (IPM), uses chemical control (selective pesticides) der, family, genus) has been proposed by several au- together with other control measures when con- thors as a way of measuring diversity, based on the sidered necessary after previous monitoring proce- existence of a strong relationship between higher dures; and Organic Agriculture (OA), does not use taxa richness and species richness (Andersen, 1995; synthetic pesticides. OA1 orchard (0.6 ha), in Vilar Gaston and Williams, 1993; Wilkie et al., 2003; Groc village, was situated at a house’s backyard, sur- et al., 2010). Therefore with a lower effort and using rounded by a non-cultivated field, pear orchards, fewer resources, a large amount of information from a vineyard and vegetable plots. It was bordered by several taxa can be obtained without identification wildflower strips and by a stone wall covered with to the species level (Gaston and Williams, 1993). In weeds. Vegetation covered the orchard’s inter-rows. fact, the higher taxa approach has already been used OA2 (1 ha), in Freiria village, was also situated at for several different organisms (e.g. Gaston and Wil- a house’s backyard. It was bordered by a road and liams, 1993; Andersen, 1995; Gaston and Blackburn, shrub hedges, surrounded by a vineyard, pear or- 1995; Martín-Piera, 2000; Ricottaet al., 2002; Biaggini chards and a non-cultivated field. It had wildflower et al., 2007; Cotes et al., 2010; Groc et al., 2010), in- strips along the field edge, inter-rows covered with cluding spiders (Borges et al., 2002; Cardoso et al., vegetation and botanically rich alleyways. IPM1 2004; Gouveia, 2004). (3.5 ha), near Caldas da Rainha city, was at a valley, However, some limitations must be considered surrounded by pine trees, eucalyptus and riparian when using this approach, because several factors vegetation, and bordered by other orchards (pear can influence the relationship between species rich- and apple). It had wildflower strips along the or- ness and higher taxa richness, namely, genealogy chard edges, botanically rich alleyways and vegeta- and status of higher taxa, choice of higher taxa rank, tion covering the inter-rows. IPM2 (3 ha), in Peral spatial scale at which the assessment occurs, areas village was bordered by a road, a small brook with with unusual patterns of speciation, and sampling riparian vegetation, wildflowers strips, hedges with period (Gaston and Williams, 1993; Andersen, 1995). trees and shrubs, and was surrounded by orchards Spiders, as predators, are beneficial arthropods in and non-cultivated fields. Vegetation covered the agroecosystems, helping in pest control, and so inter-rows. CA1 (0.1 ha), also in Peral village, was their monitoring is important (Duelli, 1997; Marc et above a valley, bordered by vegetable crops, apple al., 1999; Tavares et al., 2011). trees, weeds and piles of dried plants (reeds). On Very few studies using higher taxa surrogacy have one side was a wooden wall with nearby grassland. been conducted in agroecosystems (Duelli, 1997; Inter-rows were covered with vegetation. CA2 (4 Cotes et al., 2010); and most of them deal with nat- ha), in Vermelha village, was at a valley, bordered ural ecosystems conservation. The higher taxa ap- by other pear orchards, riparian vegetation near proach has already been used on spider’s priority a small brook, wildflowers strips, shrubs and tree conservation areas in Portugal (Borges et al., 2002; hedges and surrounded by pine trees and euca- Cardoso et al., 2004; Gouveia, 2004), but not yet in lyptus. Inter-rows were covered with vegetation. agroecosystems. Spiders were collected by ten pit-fall traps (10 cm In this work, the use of the higher taxa surrogacy diameter opening) in each orchard, placed along the approach was tested for spiders in pear orchards in intertree - strips, separated by 5 meters from each the “Oeste” region of Portugal. other. Plywood square was placed as a roof over the traps, allowing spiders to crawl under it, but avoiding the accumulation of debris or rain. Pitfall Material and Methods trapping was chosen, as it is a method directed at epigeous spiders that requires the less experience, The study took place in six pear orchards, in the effort and cost to employ (Cardoso, 2004; Cardosoet Oeste region of Portugal (Figure 1), from June to al., 2007). Additionally, it can capture almost half of November 2006 and from April to June 2007, every the spider species living in a typical Mediterranean two weeks. Two orchards of each type of production habitat and can capture a great number of individu- system were sampled: Organic Agriculture (OA1 als (Cardoso et al., 2009), which is important for the and OA2), Integrated Pest Management (IPM1 and statistical robustness of the data. IPM2) and Conventional Agriculture (CA1 and Pit-falls had 30% ethylene glycol inside. The liquid CA2). Conventional Agriculture (CA) uses broad- content was filtered with a filter paper and the solid spectrum pesticides applied rigidly at certain es- content was stored in 75% alcohol. Spiders were lat- tablished periods; Integrated Pest Management er separated from the remaining material in labora-
Tavares et al., Methodological approach for assess biodiversity 325
Figure 1 – Map of Portugal: the Oeste region, where the orchards were sampled is indicated by the square. Orchards are indi- cated by the dark shapes: OA1, OA2, IPM1, IPM2, CA1, CA2.
tory and stored in 70% alcohol and glycerine. Adult the family-species ones. When all orchards were specimens were identified to species level or morph analysed together, r2 was strong (r2>0.80) for both species. higher taxa-species relationships. When orchards For the evaluation of the potential use of family and were analysed separately, no model predominated genus as higher taxa surrogates, we used the regres- over the others, in terms of fitness. When analy- sion analysis for each orchard and for all of them to- sing all orchards together, the linear model was the gether. Three regression models were tested: a line- strongest one (highest r2) for the relationship genus- ar model (y=a+bx), an exponential model (y =a·ebx) -species and the exponential model for the rela- and a log-log model (logy=a+b·logx). Species were tionship family-species. tested as the dependent variable and the higher taxonomic categories as the independent variable. Statistical analysis was performed with STATISTI- Discussion CA 7.0 and SPSS softwares. The relationship between family-species and genus- -species was statistically significant. According to Results Martín-Piera (2000), an increase in higher taxa im- plies an increase in the number of species. In this study, 1288 adult’s spiders were collected Three regression models were tested, as proposed with the pit-fall traps. Fifteen families, 49 genera by Borges et al. (2002). According to those authors, and 64 species and morphospecies were identified the log-log model is usually the most liable model, (Table 1). because it normalizes residuals distribution and All 42 regressions (Table 2) performed were highly consequently has a more powerful statistical predic- significant (p<0.001), except the exponential model tion. However, in the present study, when orchards for the family-species relationship in CA2 (p=0.004) were analysed separately, no model revealed itself (Table 2right). Both higher taxa tested (family and stronger than the other, and when all orchards were genus) presented good relationships with species, analysed together, the linear model was the most independently of the regression models applied, adequate for the relationship genus-species and the being the genus-species relationships stronger than exponential one for the relationship family-species.
326 Revista de Ciências Agrárias, 2013, 36(3): 324-330 Table 1 – Spiders species sampled in the six orchards (total number of adults) of Organic Agriculture (OA), Integrated Pest Man- agement (IPM) and Conventional Agriculture (CA).
Family Species OA1 OA2 IPM1 IPM2 CA1 CA2 Agelenidae Malthonica lusitanica 2 0 1 0 9 0 Tegenaria atrica 1 0 0 0 0 0 Tegenaria feminea 13 2 15 4 6 0 Tegenaria montigena 0 1 3 0 1 0 Tegenaria spp. 0 0 0 0 0 1 Corinnidae Phrurolithus minimus 0 0 1 0 0 0 Liophrurillusflavitarsis 0 0 0 0 2 0 Dyctinidae Dictyna civica 1 0 0 1 1 0 Altella sp. 0 0 0 0 1 0 Dysderidae Dysdera crocata 3 2 0 1 4 0 Dysdera fuscipes 5 0 0 2 2 0 Dysdera lusitanica 0 0 2 1 0 0 Harpactea spp. 0 0 6 1 8 0 Rhode scutiventris 0 0 0 0 0 2 Gnaphosidae Drassodes lapidosus 0 0 0 1 2 0 Drassodes luteomicans 0 0 1 2 2 0 Gnaphosa alacris 1 1 0 0 2 0 Haplodrassus dalmatensis 0 2 1 1 1 0 Haplodrassus macellinus 2 0 3 0 2 0 Micaria pallipes 0 2 3 0 2 1 Trachyzelotes fuscipes 3 12 18 1 4 13 Setaphis carmeli 0 1 2 1 1 0 Zelotes aeneus 18 13 163 18 34 166 Zelotes civicus 0 0 1 0 1 0 Zelotes electus 0 0 1 0 0 0 Zelotes tenuis 7 12 29 11 22 6 Zelotes ruscinensis 0 1 0 1 2 0 Linyphiidae Canariphantes zonatus 0 0 0 0 1 0 Gongylidiellum vivum 0 0 0 1 0 0 Lessertia dentichelis 0 0 0 5 0 0 Microctenonyx subitaneus 0 0 0 5 0 0 Prinerigone vagans 0 0 0 1 0 0 Palliduphantes stygius 32 2 5 5 18 0 Pelecopsis inedita 0 0 1 0 0 0 Tenuiphantes tenuis 19 21 27 6 22 16 Linyph spp.1 1 0 0 0 0 0 Linyph spp.2 0 0 2 0 0 0 Meioneta spp. 1 0 1 0 0 0 Liocranidae Agraecina lineata 1 3 5 1 0 1 Agroeca brunnea 0 0 1 0 0 0 Lycosidae Alopecosa albofasciata 41 34 22 8 60 7 Arctosa excellens 0 0 0 3 0 0 Hogna radiata 0 1 0 0 4 0 Pardosa proxima 3 10 30 2 1 4 Pardosa hortensis 1 3 0 0 0 0 Pardosa pullata 0 0 0 0 1 0 Trochosa ruricola 0 1 0 0 0 0 Nemesiidae Nemesia bacelerei 0 0 0 0 41 0 Pisauridae Pisaura mirabilis 4 0 0 0 0 0 Salticidae Euophrys herbigrada 1 0 0 0 1 0 Euophrys sulphurea 1 0 1 0 0 0 Evarcha jucunda 0 0 1 0 0 0 Phlegra fasciata 1 0 1 0 1 0 Pseudeuophrys erratica 0 0 0 1 0 0 Talavera aequipes 0 0 0 0 1 0 Titanoecidae Nurscia albomaculata 0 1 4 1 1 1 Theridiidae Robertus arundineti 1 0 0 0 0 2 Robertus spp. 0 0 0 1 0 0 Theridion spp. 0 0 0 0 1 0 Thomisidae Ozyptila bicuspis 1 0 1 3 3 20 Ozyptila pauxilla 7 2 5 0 13 10 Xysticus kochi 0 0 0 1 0 0 Xysticus nubilus 1 0 0 0 0 0 Zodariidae Zodarion atlanticum 0 1 10 0 2 1 Total number of individuals 172 128 367 90 280 251 Number of families 12 9 12 12 14 10 Total number of species 27 22 32 29 37 15
Tavares et al., Methodological approach for assess biodiversity 327 Table 2 – LEFT: Regression analysis between the number of species (dependent variable) and the number of genera (independ- ent variable); and RIGHT: Regression analysis between the number of species (dependent variable) and the number of families (independent variable); Orchards: Organic Agriculture (OA1, OA2), Integrated Pest Management (IPM1, IPM2) and Conventional Agriculture (CA1, CA2)
The same conclusion was reached by Cardoso et al. orchards separately, the relationship family-species (2004) and Gouveia (2004). In the case of the CA2 or- presented r2 values of a wider range, while those of chard the exponential model for the family-species the relationship genus-species were less variable. relationship was not adequate. This is due to the fact Attention must be made to the fact that Andersen that although the number of families is relatively (1995), applying the higher taxa surrogacy approach high in this plot and close to the number of families for species of ants in different regions of Australia, found for all the orchards, the number of species is found that the relationship genus-species was very much lower than the one found for the other orchar- strong within regions, but varied substantially be- ds (Table 1). Thus, the exponential model will not be tween regions. However, in Portugal this has not adequate, since in this particular case, as long as we been observed: studies conducted in different re- increase the number of families, the number of spe- gions presented similar results (Borges et al, 2002; cies will not increase exponentially. CA2 revealed a Cardoso et al., 2004; Gouveia, 2004). According to very low diversity of species, being overall domina- Cardoso et al. (2004), neither geographical location ted by a single species Zelotes egregioides. Abundance nor habitat type appears to have a strong influence of a single species is characteristic of very disturbed in the higher taxa approach of spiders in Portugal. habitats (Samu, Sunderland and Szinetár, 1999). The taxon that best predicted species richness was the genus, as also concluded in other studies, name- Conclusions ly Borges et al. (2002), Cardoso et al. (2004), Gouveia (2004) and Groc et al. (2010), conducted in different This study presents additional evidence that higher ecosystems. According to Martín-Piera (2000), the ge- taxa can be used to predict species richness in a gi- nus explains a higher percentage of variance than the ven area with reasonable accuracy, being the genus family. Additionally, when analysing the different the best surrogate for species as proposed by Cardo-
328 Revista de Ciências Agrárias, 2013, 36(3): 324-330 so et al. 2004. Identification to genus level and pos- Cardoso, P.; Silva, I.; Oliveira, N.G. and Serrano, terior extrapolation through a linear model revealed A.R.M. (2004) - Higher taxa surrogates of spider to be a good methodology to be used for spider’s (Araneae) diversity and their efficiency in con- species. The relationship family-species revealed a servation. Biological Conservation, vol.117, n.4, p. lower correlation, and is not adequate for all situ- 453-459. ations such as the orchard CA2. It may be possible Cardoso, P.; Silva, I.; Oliveira, N.G. and Serrano, that disturbed habitats present a different family- A.R.M. (2007) - Seasonality of spiders (Araneae) -species relationship. in Mediterranean ecosystems and its implica- The higher taxa surrogacy approach for spiders se- tions in the optimum sampling period. Ecological ems to be an interesting method for a less expensi- Entomology, vol.32, n.5, p. 516-526. ve and time consuming biodiversity assessment in Cardoso, P.; Henriques, S.S.; Gaspar, C.; Crespo, agro ecosystems. L.C.; Carvalho, R.; Schmidt, J.B.; Sousa, P. and Szūts, T. (2009) - Species richness and composi- tion assessment of spiders in a Mediterranean Acknowledgements scrubland. Journal of Insect Conservation, vol.13, n.1, p.45-55. To Elisabete Figueiredo (ISA/UTL) and Filomena Cotes, B.; Campos, M.; Pascual, F.; García, P.A. and Magalhães (FCUL) for helping on the statistical Ruano, F. (2010) - Comparing taxonomic levels analysis and to Fernanda Amaro (INRB) for the in- of epigeal insects under different farming sys- centive. This study was financed by the Agro Pro- tems in Andalusian olive agroecosystems. Ap- gramme – Portugal (Project Agro 545). plied Soil Ecology, vol.44, n.3, p. 228-236. Duelli, P. (1997) - Biodiversity evaluation in agricul- tural landscapes: An approach at two different References scales. Agriculture, Ecosystems and Environment, vol.62, n.2-3, p. 81-91. Amaro, P. (2003) – A Protecção Integrada [em linha]. Duelli, P.; Obrist, M.K. and Schmatz, D.R. (1999) Lisboa, ISA/Press. Disponível em:
Tavares et al., Methodological approach for assess biodiversity 329 Melic, A. (Eds.) - m3m: Monografías Tercer Milenio, ment of species richness. Biological Conservation, vol.1 Zaragoza, Sociedad Entomológica Aragon- vol.107, n.1, p. 131-133. esa (SEA) e Programa Iberoamericano de Ciencia Tavares, C.A.; Gouveia, F.; Oliveira, N.G.; Crespo, y Tecnología para el Desarrollo (CYTED), p. 35-54. L.C.F.; Mateus, C. and Rebelo, M.T. (2011) - Moonen, A.-C. and Bàrberi, P. (2008) - Functional Ground-dwelling spiders (Araneae) in pear or- biodiversity: An agroecosystem approach. Ag- chards in the Oeste region of Portugal. Revista riculture, Ecosystems and Environment, vol.127, Ibérica de Aracnología, vol.19, p. 147-156. n.122, p. 7–21. Samu, F.; Sunderland, K.D. and Szinetár, C. (1999) Paoletti, M.G. (1999) - Using bioindicators based on - Scale-dependent dispersal and distribution biodiversity to assess landscape sustainability. patterns of spiders in agricultural systems: a re- Agriculture, Ecosystems and Environment, vol.74, view. The Journal of Arachnology, vol.27, n.1, p. n.1-3, p. 1-18. 325-332. Primack, R.B. (2002) –Essentials of conservation biol- Wilkie, L.; Cassis, G.; Gray, M. (2003) - A quality ogy. Sunderland, Massachusetts, Sinauer Associ- control protocol for terrestrial invertebrate biodi- ates, Inc. Publishers, 698 p. versity assessment. Biodiversity and Conservation, Ricotta, C.; Ferrari, M. and Avena, G. (2002) - Using vol.12, n.1, p. 121-146. the scaling behaviour of higher taxa for assess-
330 Revista de Ciências Agrárias, 2013, 36(3): 324-330